Introducing Artificial Intelligence into AWP Implementation

by Daniel Oliveira, CII

Advanced Work Packaging (AWP) is an execution planning Best Practice that may become the foundation of a modern production system for capital projects, as Stephen Mulva wrote in Next Generation Advanced Work Packaging. AWP is also at the core of one of CII’s overarching research programs.

I came across a Harvard Business Review article entitled Artificial Intelligence for the Real World, which describes the use of Artificial Intelligence (AI) and Machine Learning (ML) across several companies. While not specifically written with capital projects in mind, it provides insights on possible strategies, applications, and an adoption framework that could serve capital projects organizations interested in leveraging AI for AWP. Ultimately, the insights can help CII address this question: 

How can companies leverage AI to optimize AWP and achieve better business and project outcomes? 


The article indicates that companies have adopted a variety of strategies to take advantage of AI/ML. Some have adopted a bold moon-shot strategy to tackle big problems, while others have addressed low-hanging fruit (for instance, automating simple tasks first). The article recommends the latter, which is a less ambitious strategy as the path leading to better results. The low hanging fruit strategy is probably the best fit for our industry. 


The article mentions three categories of AI applications that could be extended to the AWP context: 

  • Automation of processes and tasks (cognitive automation): This could apply to the automation of tasks such as defining packages or the automation of more comprehensive AWP processes 

  • Gaining insights through data analytics: AWP data could be collected and analyzed to gain insights and optimize processes. These insights could support the optimization of startup strategies (creating a path of construction that optimizes startup); address common inefficiencies across a portfolio of projects; or identify supply chain issues before they emerge. This is where CII’s Data Warehouse can provide an ideal platform for data storage and analytics. 

  • AI applications with human engagement: These capabilities could enable planners, managers, and frontline workers to query and interact with machines to solve problems and optimize AWP implementation. For example, a Siri-like interface could provide implementation guidance. 


The article also proposes a four-step framework to integrate AI technologies: 


Transferring these recommendations to the AWP context can help CII to plan the path forward for the AWP research and frame some important questions: 

  • Which strategies can companies adopt to incorporate AI in their AWP processes?  

  • Which AWP decisions or process steps can take advantage of AI?

    • What are the key decisions in the AWP process? 

    • For each decision or process step, which AI/ML technologies are most promising?

    • For each technology, what are the data requirements?

  • Which decisions or workflow steps should be prioritized for AI/ML optimization?

    • Is there a need to redesign workflows in order to optimize the adoption of AI?

  • Which steps are bottlenecks that would benefit from automation or analytics? 

  • Which strategies can companies adopt to incorporate AI in their AWP processes?

CII will continue to support members’ efforts to understand, implement, and develop AWP in their organizations. Our research program will tackle these challenges and ultimately promote step changes in the way AWP improves business and project outcomes.


Artificial Intelligence for the Real World by Thomas H. Davenport and Rajeev Ronanki 

Harvard Business Review. “HBR’s 10 Must Reads on AI, Analytics, and the New Machine Age” (with bonus article “Why Every Company Needs an Augmented Reality Strategy” by Michael E. Porter and James E. Heppelmann) (Kindle Locations 101-103). Harvard Business Review Press. Kindle Edition.

Date posted: February 19, 2020